R/loglikcushcov.R

In CUB: A Class of Mixture Models for Ordinal Data

#' @title Log-likelihood function for a CUSH model with covariates
#' @aliases loglikcushcov
#' @description Compute the log-likelihood function for a CUSH model 
#' with covariates for the given ordinal responses.
#' @usage loglikcushcov(m, ordinal, X, omega, shelter)
#' @param m Number of ordinal categories
#' @param ordinal Vector of ordinal responses
#' @param X Matrix of selected covariates for explaining the shelter parameter
#' @param omega Vector of parameters for explaining the shelter effect, 
#' with length equal to NCOL(X)+1 to account for an intercept term (first entry of omega)
#' @param shelter Category corresponding to the shelter choice
#' @keywords internal
#' 


loglikcushcov <-function(m,ordinal,X,omega,shelter){
  if (is.factor(ordinal)){
    ordinal<-unclass(ordinal)
  }
  X<-as.matrix(X)
  deltavett<-logis(X,omega)
  dummy<-ifelse(ordinal==shelter,1,0)
  probi<-deltavett*(dummy-1/m)+1/m
  return(sum(log(probi)))
}